Realizing%20Programmable%20Matter - PowerPoint PPT Presentation

About This Presentation
Title:

Realizing%20Programmable%20Matter

Description:

Realizing Programmable Matter. Seth Copen Goldstein and Peter Lee ... Gerry Sussman (MIT) Bill Swartout (ICT) David Tarditi (Microsoft) Bob Tulis (SAIC) ... – PowerPoint PPT presentation

Number of Views:46
Avg rating:3.0/5.0
Slides: 50
Provided by: sethcopen
Learn more at: https://www.cs.cmu.edu
Category:

less

Transcript and Presenter's Notes

Title: Realizing%20Programmable%20Matter


1
Realizing Programmable Matter
  • Seth Copen Goldstein and Peter Lee
  • DARPA POCs Jonathan Smith and Tom Wagner (alt.)

2006 DARPA ISAT Study
2
What is Programmable Matter?
ETC, 2006
3
What is Programmable Matter?
  • A programmable material
  • with actuation and sensing
  • that can morph into shapes under software
    control
  • and in reaction to external stimuli

4
Using Programmable Matter
Protenna
Time
5
Key questions
  • Can we really make programmable matter?
  • If we make it, can we write useful programs for
    it?
  • Are there reasons to do this now?
  • What are potential applications?

6
Can we really make programmable matter?
7
Starting points
UCB
TI
Klavins/UW
Stoddart
UCLA
Sciam
Yim/Parc
MIT
Stoy/USD
Storrs-Hall/Rutgers
MIT
8
A fundamental goal Scaling
  • Consider applications that involve rendering
    macroscale objects
  • High fidelity rendering implies
  • sub-millimeter-scale units (voxels)
  • massive numbers of units
  • Units must be inexpensive
  • mass-produced
  • largely homogeneous
  • simple, possibly no moving parts

9
A fundamental goal Scaling
Modular robotics
Stoy
Focus of this talk micron (MEMS) scale
Nano/chemistry
Stoddard
10
A potential approach
  • How to form 3D from a 2D process?
  • begin with foundry CMOS on SOI

11
A potential approach
  • How to form 3D from a 2D process?
  • begin with foundry CMOS on SOI
  • pattern a flowerthat includes structure and
    circuits

12
A potential approach
  • How to form 3D from a 2D process?
  • begin with foundry CMOS on SOI
  • pattern a flowerthat includes structure and
    circuits
  • lift off silicon layer
  • flexible
  • harness stress to form a sphere

13
A sanity check
Computation Capability 8086 Processor with 256KB
memory SOI-CMOS 90 nm process with gt 2M
transistors.
1 mm diameter sphere
Mass lt 1 mg
Electrostatic Actuators 5 body lengths / sec
Communication Capacitors
Power Storage Supercapacitor stores enough energy
to execute over 200 million instructions or move
2 million body lengths
Power distribution Transmission of energy
packets using capacitive coupling fills
reservoir in lt 1?s.
14
Additional challenges
  • We investigated concepts in integration of
  • adhesion mechanisms
  • power distribution
  • energy storage
  • communication
  • heat management

15
Major milestones (hardware)
time

communication and localization for sensing of (interior and exterior) shapes dynamic localization and active adhesion for a digital clay control for simple coordinated actuation integration for coordinated sensing and actuation macro-scale rendering and dynamic shape shifting general distributed programming model
device integration network initial power programmable adhesion power and heat management actuation sensor integration display biomemetic and/or chemical adhesion
FPF
functionality
hardware requirements
16
Can we really make programmable matter?
Probably. But then can we program programmable
matter?
17
Programming large machines
  • Concepts in parallel, distributed, and
    high-performance computing
  • Can scale to thousands of nodes for
    embarrassingly parallel applications,
  • with known, regular interconnect
  • But how do we program millions of mobile,
    interacting devices?

18
Algorithms vs control
  • Our study considered the programming problem at
    two levels
  • Programming the Ensemble How does one think
    about coordination of millions of elements?
  • Programming the Unit What is the programming
    model for a (single) element?

19
Physical rendering
  • To simplify our approach, we focus exclusively on
    physical rendering
  • How to coordinate the movement of the units to
    form a desired physical shape
  • Today Motion planning
  • But with a large number of units, central motion
    planning is not tractable
  • A stochastic approach appears to be necessary

20
Potential Approaches
Lipson
Nagpal
Klavins
DeRosa
Stoy
21
Potential Approaches
Lipson
Nagpal
Klavins
DeRosa
Stoy
22
Hole flow methods
DeRosa
23
Rendering
  • Conclusion For rendering a stochastic approach
    appears to have several advantages
  • exploits large numbers
  • requires no central planning
  • simple specification
  • scale-independent
  • robust to failures in individual elements

24
Global Behavior from local rules
  • Concise specifications
  • Embarrassingly parallel
  • Examples
  • Amorphous computing Nagpal
  • Graph grammars Klavins
  • Programming work Kod.
  • CAGradients Stoy
  • Hole motion DeRosa
  • Boyd model Boyd
  • Turing stripes
  • Goal Compile Global specification into unit
    rules
  • Predict global behavior from set of unit rules

25
Major Software milestones
time

Localization Power routing Communication Unit control External sensing Robustness to lattice faults Locomotion Failing units Distributed inference global behavior from local rules Thermodynamics of programming Planning General distributed programming models
FPF
functionality
HW
Simulate unit to unit motion To feed hw unit design Simulate PM dynamics Verify hdware sensor reqs Simulate large scale env. interaction Robust to hdware faults
Simluation
SW
26
TowardsThermodynamics of Programming
27
Why should DARPA invest in programmable
matter? Would a soldier use an antenna made out
of PM?
28
Versatility and efficiency
Versatility is great, but has a cost
  • For some instances, PM would be
  • lower performance
  • complicated
  • expensive
  • FPGAs are also
  • slow
  • large
  • power hungry
  • and the fastest-growing segment of the silicon
    market

29
Field programmability for the physical world
Benefit
Capability
Copes easily with low volumes typical in military applications Rapid production with lowered factory retooling costs
Fast response to military needs Situation-specific hardware on demand
Easy upgrades in the field Adapt equipment to lessons learned in the field
One device for many purposes, combinable with those carried by others Reduce SWAP and logistics load
Change and create equipment for new conditions Specialized equipment for unpredictable situations
Production volume
Time to market
Upgrades
Functionality
Adaptability
30
Furthermore
  • Programmable Matter is
  • scalable and separable
  • PM carried by many soldiers can be combined for
    larger objects
  • computational / reactive
  • reconfiguration can be dynamic, reactive to
    environment
  • Valuable in situations where time and
    distance matter
  • space, ships, embassies, convoys,
  • quick fixes, decoys, improvisation

31
Uses in the field
  • PM in the field takes on useful shapes
  • physical display / sand table
  • specialized antennas
  • field-programmable mold
  • shape dirt and elastomericcross-linked polymer
    intobullet-proof objects
  • mold customized shaped charges
  • 3D fax
  • In CONUS, needed object is designed or
    PM-captured, then sent to the field

32
Understanding Complexity
Nanotechnology is more than just small
  • Future applications of nanotechnology atthe
    macroscale require study of Systems
    Nanotechnology
  • Programmable matter is a key enabler for studying
    large complex systems

The science and technology of manipulating
massive numbers of nanoscale components
33
Heilmeier questions
  • What are we trying to do?
  • Build a programmable material that is able to
    morph into shapes, under software control and in
    reaction to external stimuli. Bring power of
    programming to the physical world.
  • How is it done today? What are the limitations of
    current practice?
  • Preplanning, prepositioning, and many specialized
    objects. This means big loads and lack of
    flexibility to handle unforeseen needs.
  • What is new in our approach why do we think it
    can succeed?
  • Potential designs indicate feasibility of the
    hardware. Physical rendering is a sweet spot
    that is tractable, software-wise.
  • Assuming we are successful, what difference will
    it make?
  • New capabilities in low-volume manufacturing and
    3D displays. Antennas may achieve radical
    improvements. New programming models for and
    understanding of large-scale systems.
  • How long will it take? How much will it cost?
  • Basic units can be built in the near term.
    Integration of adhesion, sensing, locomotion
    several years later, leading to initial
    deployable applications in the 5-10 year time
    frame.

34
Conclusions
  • Manufacturing PM elements poses challenges, but
    appears to be feasible and may lead to new 3D
    concepts
  • Software for PM applications, while raising
    significant questions, appears algorithmically
    feasible for physical rendering but still
    requires breakthroughs in distributed computing
  • Application domain of rendering can form
    springboard for advances in models and languages
    for massively distributed programming of reality
  • There are leap-ahead military applications, in
    both longer and shorter time frames

35
Participants
  • Tayo Akinwande (MIT)
  • Lorenzo Alvisi (UT-Austin)
  • Michael Biercuk (BAH)
  • Jason Campbell (Intel)
  • Brad Chamberlain (Washington)
  • Bob Colwell (Intel)
  • Andre DeHon (UPenn)
  • John Evans (DARPA)
  • Gary Fedder (CMU)
  • Alan Fenn (MIT-LL)
  • Stephanie Forrest (UNM)
  • Seth Goldstein (CMU)
  • James Heath (CalTech)
  • Maurice Herlihy (Brown)
  • Peter Kind (IDA)
  • Eric Klavins (Washington)

Tom Knight (MIT) Dan Koditschek (UPenn) Peter
Lee (CMU) Pat Lincoln (SRI) Hod Lipson
(Cornell) Bill Mark (USC-ISI) Andrew Myers
(Cornell) Radhika Nagpal (Harvard) Karen Olson
(IDA) George Pappas (UPenn) Keith Kotay
(MIT) Zach Lemnios (MIT-LL) Kathy McDonald
(SOCOM) Dan Radack (DARPA) Rob Reid (AFRL) John
Reif (Duke)
Daniela Rus (MIT) Vijay Saraswat (IBM) Metin
Sitti (CMU) Jonathan Smith (DARPA) Dan Stancil
(CMU) Guy Steele (Sun) Allan Steinhardt
(BBN) Gerry Sussman (MIT) Bill Swartout
(ICT) David Tarditi (Microsoft) Bob Tulis
(SAIC) Tom Wagner (DARPA) Janet Ward
(RDECOM) Mark Yim (UPenn) Marc Zissman (MIT-LL)
ISAT member
36
BACKUP SLIDES FOLLOW
37
Calculating the voltage
38
Relative Locomotion on mm scale
  • Locomotion Constraints
  • Modules motions are discrete on lattice
  • (e.g. simple cubic, body-centered-cubic).
  • Face detaches
  • Module moves along simple 1DOF path
  • New face-face latches
  • Constraints
  • Modules move only self (or neighbor)
  • Assume modules remain connected (for power)
  • Worst case forces lift one module against
    gravity.
  • Actuation Technology
  • Electrostatic (baseline)
  • Electromagnetic
  • Hydrophillic forces
  • External actuation

2 rhombic dodecahedrons n BCC lattice
2 Spheres on cubic lattice (one moving)
39
Embedded computers
  • Embedded processors dominate
  • 300 million PCs and servers
  • 9000 million embedded!

40
Costs of micro-scale device
  • Module 1mm x 1mm x 1mm MEMS (silicon)
  • Silicon cost 1/sq inch
  • 2003 Revenue 5.7billion / 4.78 billion sq inch
    silicon
  • 200 / 12 diam, 30 /8 diam wafers
  • 100um-2000um thick (choose 1mm)
  • Assume processing costs 9/sq inch
  • Modules cost 1.6
  • Average person weighs 65 Kg -gt 65,000 cm3
  • Assume density of water (1kg 1000 cm3 )
  • 65,000,000 modules
  • 1000 modules per cm3
  • Cost 1,007,502
  • More realistic, rendering of the shell 1,500,000
    modules 24,000

41
Robustness
  • Large distributed systems (6 nines for each
    unit ? less than 1 nine for the ensemble)
  • Acting in the real world
  • Environmental uncertainty
  • Parametric uncertainty
  • Harsher than the machine room (plain old
    faults/defects)
  • Known problem in robotics and distributed systems
  • Current approaches dont scale or are not
    integrated
  • Make Uncertainty Tolerance first class

42
Embrace Stochastic Approaches
  • Need reliable (but not exact) outcomes from
    unreliable components and information
  • Information Based Complexity shows
  • when information is
  • costly,
  • tainted,
  • partial
  • Worst-case error bounds require exp-time. with
    high-probability error bounds require
    poly-time!
  • Programmable Matter has
  • costly communication,
  • noisy sensors,
  • no one unit has the whole picture
  • Emerging paradigms for unit control
  • hybrid vs. discrete computation
  • converges toward acceptable result

43
Topological Approaches to Unit Control and
Composition
Deform
State Space view
¼
physical problem instance
topological model of physical problem instance
point attractor basin
¼
Composition Operator
¼
topological model of point attractor basin
sequential composition of point attractor basins
44
Software trajectory
  • There is path
  • Rendering is sweet spot
  • Research directions
  • Make uncertainty tolerance first class
  • Embrace stochastic behavior
  • Outcome
  • Develop a thermodynamics of programming languages
    which will lead to
  • Compiling specification into unit rules
  • Predict global behavior from local rules

45
A Proposed unit of PM
1 mm diameter sphere
Processor 1x 8086s with 256KB memory Formed from
CMOS imbedded in glass layers. Using 50 of the
surface area provides over 500K transistors with
a 90 nm CMOS process.
Surface Area 3.14 sq. mm. Volume 0.52 cu.
mm Mass lt 1 mg
Electrostatic Actuators/ Communication
Capacitors Formed using top level CMOS metal
layer, can be located above processing elements
Power distribution Uses metal lines fabricated
using CMOS and enclosed in glass.
Power Storage Super cap integrated in the
interior of the sphere/polyhedron 1J per cubic cm
equates to 0.26 mJ
46
Feasibility
  • Area 1mm diameter, ? mm2
  • 50 for circuits
  • 90nm 2M transistors
  • 180nm 500K transistors
  • Computation Memory
  • 8086 (30K Ts) 1 Mip
  • Program size 64K
  • Total RAM 256K
  • Energy
  • supercap 50 volume .26mJ
  • 1pJ/instruction
  • 70 pJ/body length
  • mass (density of glass)
  • .7mg
  • Locomotion by electrostatic coupling
  • lt400V generates 80 ?N
  • lt50 ms for 180 degree rotation
  • Energy transfer by cap coupling
  • Deliver .026mJ in .24ns
  • Fill reservoir in 24ns
  • Adhesion
  • Fast ES several units in worst case
  • Others surface tension, covalent bonds
  • Cost
  • 9/in2
  • Unit 0.016

47
Field programmable concepts
  • From the Natick Soldier Systems Center and
    Special Operations
  • precisely shaped explosive charges
  • mortar base plate
  • gun magazines
  • PJ equipment
  • field radio
  • one-handed bandages

48
What is Nanotechnology?
http//www.powersof10.com/
49
A sanity check
1 mm diameter sphere
Processor 1x 8086s with 256KB memory Formed from
CMOS imbedded in glass layers. Using 50 of the
surface area provides over 2M transistors with a
90 nm CMOS process.
Surface Area 3.14 sq. mm. Volume 0.52 cu.
mm Mass lt 1 mg
Electrostatic Actuators/ Communication
Capacitors Formed using top level CMOS metal
layer, can be located above processing elements
Power Storage A supercap integrated in the
interior of the sphere/polyhedron Stores enough
energy to execute over 200 million instructions
or move 2 million body lengths
Power distribution Unit-unit via capacitive
coupling and transmission of energy packets.
Interior routing to central storage capacitor.
Write a Comment
User Comments (0)
About PowerShow.com